Skip to content
Special Feature · U.S. Economy

Recession Dashboard

Eleven indicators economists watch for the next U.S. recession, from the yield curve to jobless claims, each with a plain verdict, the rule behind it and how often it has flashed before or around the start of past recessions. Updated every day from official data.

0 of 11
indicators flashing a recession warning, 2 more on watch · The indicators are mostly clear. · updated November 11, 2026
0.68 pts10y-3m spread, Nov 10, 2026 · Watch
0.00 ptsSahm rule, October 2026 · Clear
−1.1%Jobless claims, Oct 31, 2026 · Clear
+137kPayroll jobs, October 2026 · Clear
3.4%Real GDP, Q3 2026 · Clear

01The Scorecard

Each indicator gets one of three verdicts from a fixed rule, shown with it: Flashing means it is at a level that has usually come with or before recessions, Watch means it is close (for the yield curves, that an inversion of three months or more ended within the past 18 months), and Clear means it is not. No single indicator is reliable on its own; the number flashing together matters more. That number is a plain count, not a recession probability: every indicator counts the same, and several move together (the two yield curves, and the job-market measures), so it is not calibrated against history.

IndicatorLatestVerdictRecord since data began
Yield curve: 10-year minus 3-month TreasuryNov 10, 20260.68 ptsWatch4 of 4 recessions (4 before the start, median 12 months ahead), 1 false alarm
Yield curve: 10-year minus 2-year TreasuryNov 10, 20260.25 ptsClear5 of 6 recessions (5 before the start, median 18 months ahead), 4 false alarms
Sahm rule recession indicatorOctober 20260.00 ptsClear8 of 8 recessions (2 before the start, median 9.5 months ahead), 2 false alarms
Initial jobless claims, 4-week average, change from its low of the previous 52 weeksOct 31, 2026−1.1%Clear8 of 8 recessions (4 before the start, median 10 months ahead), 3 false alarms
Payroll jobs, 3-month average monthly changeOctober 2026+137kClear7 of 8 recessions (1 before the start, median 15 months ahead), 2 false alarms
Real GDP growth, annualized, latest quarterQ3 20263.4%Clear8 of 8 recessions (5 before the start, median 3 months ahead), 4 false alarms
Industrial production, change from a year earlierSeptember 20262.2%Clear6 of 8 recessions (2 before the start, median 10 months ahead), 1 false alarm
Building permits, 3-month average, change from a year earlierSeptember 20261.0%Clear6 of 8 recessions (4 before the start, median 9.5 months ahead), 3 false alarms
Temporary help jobs, change from a year earlierOctober 20263.0%Clear3 of 3 recessions (1 before the start, median 1 month ahead), 1 false alarm
Real personal income excluding government transfers, change from a year earlierSeptember 20260.5%Watch7 of 8 recessions (1 before the start, median 18 months ahead), 4 false alarms
Real retail sales, 3-month average, change from a year earlierSeptember 20262.1%Clear2 of 3 recessions (0 before the start), 1 false alarm

Record: how many recessions that began at least 18 months after the indicator's data starts were preceded or accompanied by a flashing signal, counting the 18 months before each recession began through 6 months after, and how many flashing spells had no recession under way or beginning within 18 months of the spell's start (spells that began in the last 18 months are not counted yet). Because the window runs to 6 months after a recession began, a signal that first appeared once the recession was already under way still counts as a hit, so the record measures early or same-time detection, not only advance warning; the number in parentheses is how many of those recessions saw the signal before the starting month, and the median lead among them. The record uses today's revised data, not the figures available at the time, except the Sahm rule, which uses real-time data. Computed by DaveManuel.com from the data; recession dates are the National Bureau of Economic Research's.

02Each Indicator's History


Grey bands are recessions as dated by the NBER. The dashed line is the level at which the indicator is counted as flashing. Monthly values (GDP is quarterly) from the start of each series' data; daily and weekly series are averaged by month.

The Recession Clock: the Yield Curve vs the Sahm Rule, Month by Month

03What Each Indicator Measures

Yield curve: 10-year minus 3-month Treasury · Watch, 0.68 pts (Nov 10, 2026)

Flashing when the spread is below zero (inverted). On watch for 18 months after an inversion of three months or more ends: since the late 1960s, recessions have begun either while the curve was still inverted or within about a year of it turning positive again. Last inverted for 3 months or more until August 2025.

When short-term Treasury rates rise above long-term ones, bond investors are betting that the Federal Reserve will have to cut rates, usually because the economy is weakening. The 10-year minus 3-month spread is the version the Federal Reserve Bank of New York uses in its recession model. Inversions of three months or more in the data: Jun 1989 to Aug 1989 (next recession: peak Jul 1990); Jul 2000 to Jan 2001 (next recession: peak Mar 2001); Aug 2006 to May 2007 (next recession: peak Dec 2007); May 2019 to Sep 2019 (next recession: peak Feb 2020); Nov 2022 to Nov 2024 (no recession as of Nov 2026); Jun 2025 to Aug 2025 (no recession as of Nov 2026).

Yield curve: 10-year minus 2-year Treasury · Clear, 0.25 pts (Nov 10, 2026)

Flashing when the spread is below zero (inverted). On watch for 18 months after an inversion of three months or more ends: since the late 1960s, recessions have begun either while the curve was still inverted or within about a year of it turning positive again. Last inverted for 3 months or more until August 2024.

Sahm rule recession indicator · Clear, 0.00 pts (October 2026)

Flashing at 0.50 or above: the three-month average unemployment rate is half a point or more above its lowest three-month average of the previous 12 months. On watch from 0.30.

Named for economist Claudia Sahm, who proposed it in 2019 as a trigger for automatic stimulus. It looks at the unemployment rate, the most widely followed sign of a weakening job market. The version used here is FRED's real-time series, which uses the unemployment data as it was available at the time, not as later revised.

Initial jobless claims, 4-week average, change from its low of the previous 52 weeks · Clear, −1.1% (Oct 31, 2026)

Flashing when the four-week average of new claims is 20% or more above its lowest point of the previous 52 weeks (the latest week is not part of that low, so a negative figure means a new 52-week low); on watch from 10%. Latest four-week average: 194,750.

New claims for unemployment insurance, published weekly (usually on Thursday), are the most up-to-date measure of layoffs. A sustained rise from the low is an early sign that employers are cutting. The low is the lowest four-week average of the 52 weeks before the latest week, so a figure below zero means claims are at a new 52-week low.

Payroll jobs, 3-month average monthly change · Clear, +137k (October 2026)

Flashing when the economy loses jobs on a three-month average; on watch below 50,000 a month. Figures in thousands of jobs.

Real GDP growth, annualized, latest quarter · Clear, 3.4% (Q3 2026)

Flashing when the economy shrank in the latest quarter (two shrinking quarters in a row is the popular definition of a recession, though not the official one); on watch below 1%. The quarter before: 2.2%.

Industrial production, change from a year earlier · Clear, 2.2% (September 2026)

Flashing when factory, mining and utility output is down 2% or more from a year earlier; on watch when it is down at all.

Building permits, 3-month average, change from a year earlier · Clear, 1.0% (September 2026)

Flashing when permits for new homes are down 20% or more from a year earlier (three-month average); on watch from 10%.

Temporary help jobs, change from a year earlier · Clear, 3.0% (October 2026)

Flashing when temporary staffing jobs, often the first cut, are down 5% or more from a year earlier; on watch when they are down at all.

Real personal income excluding government transfers, change from a year earlier · Watch, 0.5% (September 2026)

Flashing when income from work and investments, after inflation, is no higher than a year earlier; on watch below 1% growth. One of the measures the NBER uses to date recessions.

The National Bureau of Economic Research, which dates U.S. recessions, has in recent decades put the most weight on payroll jobs and real personal income excluding transfers, alongside spending, sales and industrial production.

Real retail sales, 3-month average, change from a year earlier · Clear, 2.1% (September 2026)

Flashing when retail and food service sales, after inflation, are down 2% or more from a year earlier; on watch when they are down at all.

04U.S. Recessions Since 1960

PeakTroughLength
February 2020April 20202 months
December 2007June 200918 months
March 2001November 20018 months
July 1990March 19918 months
July 1981November 198216 months
January 1980July 19806 months
November 1973March 197516 months
December 1969November 197011 months
April 1960February 196110 months

Business cycle peaks and troughs as dated by the NBER's Business Cycle Dating Committee; a recession runs from the month after the peak to the trough, and its length is the number of months from peak to trough. The NBER announces these dates months, sometimes more than a year, after the fact, so a recession can be under way before it is official.

05Questions People Ask

Who decides when the U.S. is in a recession?

The Business Cycle Dating Committee of the National Bureau of Economic Research, a private nonprofit. It defines a recession as a significant decline in economic activity that is spread across the economy and lasts more than a few months, and it looks at jobs, income, spending, sales and production rather than at GDP alone. It names the start and end months only after the data confirms them.

Is a recession two quarters of falling GDP?

That is a popular rule of thumb, not the official definition. Most recessions do include two quarters of falling GDP in a row, but not all: the 1960-61 and 2001 recessions never had two in a row. In 2022 the first estimates showed two straight quarters of decline and the NBER did not call a recession; later revisions turned the second quarter positive.

Which indicator is most reliable?

Each has a different strength. The yield curve gives the longest warning, often a year or more, but its timing varies and it has given false signals. The Sahm rule and jobless claims react later but, historically, closer to the start of a recession. That is why this page counts how many are flashing at once.

How often is this page updated?

Every day. Each indicator updates when its source publishes: Treasury yields daily, jobless claims weekly, jobs, production, income and sales monthly, and GDP quarterly. Each reading shows the period it covers.

06Sources

Related: Where Your Tax Dollars Went Yesterday · U.S. National Debt Clock · The Stock Market After Midterm Elections

Corrections and Updates

DateWhat changed
October 25, 2026Track record now shows how many hits came before each recession began and the median lead; jobless-claims indicator redefined as change from the low of the previous 52 weeks (a negative figure is a new 52-week low).

Every correction to a published figure on this site is logged on Sources and Methods.

Cite This Page

This page is free to cite and link. Its figures come from FRED (Federal Reserve Bank of St. Louis), from BLS, BEA, Census and Federal Reserve data and are kept current, so please include the date you accessed it.

FormatCitationCopy
PlainDave Manuel, “Recession Dashboard,” DaveManuel.com, updated November 12, 2026, https://www.davemanuel.com/recession-dashboard.php (accessed November 12, 2026).Copy
APAManuel, D. (2026). Recession Dashboard. DaveManuel.com. Retrieved November 12, 2026, from https://www.davemanuel.com/recession-dashboard.phpCopy
ChicagoManuel, Dave. “Recession Dashboard.” DaveManuel.com. Accessed November 12, 2026. https://www.davemanuel.com/recession-dashboard.php.Copy
Wikipedia{{cite web |url=https://www.davemanuel.com/recession-dashboard.php |title=Recession Dashboard |last=Manuel |first=Dave |website=DaveManuel.com |access-date=2026-11-12}}Copy
Link (HTML)<a href="https://www.davemanuel.com/recession-dashboard.php">Recession Dashboard (DaveManuel.com)</a>Copy

Source data: FRED (Federal Reserve Bank of St. Louis), from BLS, BEA, Census and Federal Reserve data. Updates: Daily. Page last updated November 12, 2026. How we build our data pages.